On Sun, Jan 23, 2022 at 11:31 AM Noel Chiappa via cctalk
<cctalk at classiccmp.org> wrote:
  See:
   
https://walden-family.com/impcode/imp-code.pdf
 Someone's already done the specialist OCR to deal with faded program listings. 
Neat. Though all the complex character recognition part of that work
is now like 15-20 lines of Python code (using either Keras or
PyTorch).